foraging effort of juvenile steller sea lions eumetopias jubatus

ENDANGERED SPECIES RESEARCHEndang Species Res

Vol. 10: 145–158doi: 10.3354/esr00260

Published online March 3, 2010

INTRODUCTION

Over the past 30 yr, populations of Steller sea lionsEumetopias jubatus in western Alaska (USA) havedeclined by 80% (Braham et al. 1980, Loughlin 1998,Fritz et al. 2008). Decreased juvenile survival andreproductive rates have been implicated as proximatefactors for the decline of the endangered, western dis-tinct population segment (wDPS) of Steller sea lions

(York 1994, Merrick 1995, York et al. 1996, Holmes &York 2003, Holmes et al. 2007). Nutritional stressresulting from changes in distribution, abundance, orquality of prey due to commercial fisheries and large-scale oceanographic changes is among the ultimatefactors proposed for explaining the decline (Merrick1995, Loughlin 1998, Loughlin & York 2000, DeMaster& Atkinson 2002). Short-term environmental variabil-ity and local environmental perturbations have also

© Inter-Research 2010 · www.int-res.com*Email: [email protected]

Foraging effort of juvenile Steller sea lions Eumetopias jubatus with respect to heterogeneity

of sea surface temperature

Michelle E. Lander1, 2,*, Thomas R. Loughlin1, Miles G. Logsdon3, Glenn R. VanBlaricom2, Brian S. Fadely1

1National Marine Mammal Laboratory, NMFS, 7600 Sand Point Way N.E., Seattle, Washington 98115, USA2Washington Cooperative Fish and Wildlife Research Unit, School of Aquatic and Fishery Sciences, Box 355020,

University of Washington, Seattle, Washington 98195, USA3School of Oceanography, Box 357940, University of Washington, Seattle, Washington 98195, USA

ABSTRACT: Among many other factors, the decline of the western distinct population segment ofSteller sea lions Eumetopias jubatus in Alaska (USA) has been attributed to changes in the distribu-tion or abundance of prey due to the cumulative effects of fisheries and large-scale climate change.However, the depletion of localized prey resources due to small-scale environmental variability andperturbations may be impeding recovery, resulting in the need to understand how the environmentcurrently affects this species on smaller spatial and temporal scales. The objective of this study, there-fore, was to assess how Steller sea lions respond to changes in localized environmental features.Satellite-relayed data loggers were deployed on juvenile Steller sea lions (n = 24) from July 2002 toMay 2004 in the Aleutian Islands and Gulf of Alaska. Weekly indices of foraging effort (mean andmaximum trip duration, diving activity) of Steller sea lions were examined with respect to corre-sponding patterns of sea surface temperature (SST) data obtained from the moderate resolutionimaging spectroradiometer. An assortment of landscape metrics was used to characterize the hetero-geneity of frontal features derived from SST gradients because it has been suggested that Steller sealions depend on prey patches associated with these features. Multivariate analyses indicated thatfractal dimension and patch density of frontal features were significant factors for predicting differ-ent aspects of foraging effort (p < 0.05; n = 6 models). Overall, results suggested that aggregatedfrontal features associated with small-scale temperature gradients were probably conducive to forag-ing effort of Steller sea lions, but additional mechanisms should be investigated further.

KEY WORDS: Eumetopias jubatus · Environmental heterogeneity · Foraging effort · Fractaldimension · Frontal features · Gradients · Sea surface temperature · Steller sea lion

Resale or republication not permitted without written consent of the publisher

Contribution to the Theme Section ‘Biologging technologies: new tools for conservation’ OPENPEN ACCESSCCESS

Endang Species Res 10: 145–158

been suggested as hypotheses for population decline(Pascual & Adkinson 1994, Merrick 1995, Benson &Trites 2002) or lack of recovery (Fritz & Hinckley 2005,Atkinson et al. 2008, NMFS 2008), resulting in theneed to understand how the environment currentlyaffects Steller sea lions at finer scales in local coastalareas (Trites et al. 2007).

Steller sea lions heavily utilize nearshore habitats(Merrick & Loughlin 1997, Loughlin et al. 2003, Raum-Suryan et al. 2004) by adopting a strategy of centralplace and multiple central place foraging to cope withthe spatial and temporal distribution of localized preyresources (Raum-Suryan et al. 2004). Although the for-aging behavior of juvenile Steller sea lions has beenfairly well detailed (Merrick & Loughlin 1997, Loughlinet al. 2003, Raum-Suryan et al. 2004, Fadely et al. 2005,Pitcher et al. 2005, Call et al. 2007), few studies haveassessed the effects of environmental features on for-aging effort. Loughlin et al. (2003) satellite-tagged 25juvenile Steller sea lions from 1994 to 2000 in Alaskaand Washington and concluded that Steller sea lionshave the foraging flexibility to take advantage of pre-dictable behavioral traits of prey species and localizedoceanographic conditions that enhance prey concen-trations. Others have suggested that Steller sea lionsare constrained by prey persistence (Gende & Sigler2006) and changes in prey availability resulting fromseasonal variability (Merrick & Loughlin 1997, Womble& Sigler 2006), and that they depend on the presenceof large, dense prey patches associated with nearshoretemperature gradients (Sinclair & Zeppelin 2002).Hence, further analyses of telemetry data are neededto understand relationships of foraging behaviors withenvironmental variability and spatial patterns ofoceanographic structure, which ultimately affect thedistribution or abundance of prey (Loughlin et al.2003). For example, the oceanographic structure ofwater temperature likely influences prey distribution,which in turn affects foraging behavior and possiblyfecundity or mortality of Steller sea lions (Pascual &Adkinson 1994).

Spatial and temporal heterogeneity of the environ-ment have been empirically and conceptually chal-lenging to ecologists because complex environmentsare difficult to describe quantitatively. Furthermore,ecologists have become increasingly aware of theimportance of examining ecological processes at scalesrelevant to the organism and process under study(Turner et al. 1989, Wiens 1989). This is especiallyimportant in the marine environment, which is a highlydynamic system. However, with advances in geo-graphic information systems (GIS) and remote sensingtechniques, the spatial heterogeneity of the marineenvironment can be characterized over time by anassortment of patches and gradients (White & Brown

2003), which are considered the structural and func-tional components of landscapes (Cadenasso et al.2003). Spatial patterns of ocean structure and hetero-geneity can be quantified through time using anassortment of landscape metrics, which focus on theabundance (i.e. composition) and spatial arrangementor complexity (i.e. configuration) of patches (Gustafson1998, McGarigal et al. 2002). The quantification ofenvironmental heterogeneity using these metrics hasbecome a common practice in the advancing field oflandscape ecology, which is the study of landscapepatterns, ecological processes that influence patterns,and effects of patterns on population persistence andanimal movement (Hargis et al. 1997, Fahrig & Nuttle2003, Lovett et al. 2003). Although landscape ecologyhas traditionally focused on ecological processes andspatial patterns in terrestrial ecosystems, the principlesof this discipline can also be applied to aquatic systems(Wiens 2002).

If environmental heterogeneity is in part responsiblefor changing the abundance or distribution of prey forSteller sea lions on a localized scale, then changes inenvironmental features should be reflected in patternsof sea lion foraging behavior. The objective of thisstudy, therefore, was to assess how Steller sea lionsrespond to changes in heterogeneity of the environ-ment at spatial scales relevant to individual sea lionperception. More specifically, the foraging effort ofSteller sea lions was examined with respect to anassortment of landscape metrics that were used tocharacterize spatial patterns of sea surface tempera-ture (SST) gradients. We chose to work with SST as anenvironmental indicator because it has been hypothe-sized to affect Steller sea lions on multiple scales (Sin-clair & Zeppelin 2002, Trites et al. 2007) and has beenlinked to food habits and population dynamics atregional scales (Call & Loughlin 2005, Lander et al.2009). Additionally, SST allows for differentiatingwater mass structure (Boyd et al. 2001) and can beused to derive oceanographic frontal features, whichwe elaborate on in the methods.

MATERIALS AND METHODS

Juvenile Steller sea lions were opportunisticallycaptured at rookeries or haulout sites within the Gulfof Alaska and Aleutian Islands using hoop nets ordive captures (McAllister et al. 2001; Table 1, Fig. 1).Animals were either sedated with valium (1.1 to2.0 ml) and manually restrained or anesthetized(Heath et al. 1997). Sea lions were weighed to thenearest 0.5 kg, and standard length and axial girthwere measured to the nearest 1.0 cm. Measurementsof tooth size (upper canine), body size, and time of

146

Lander et al.: Steller sea lions and SST heterogeneity

year were used to infer ages of all sea lions (King etal. 2007). Satellite relayed data loggers (SRDLs; SeaMammal Research Unit [SMRU], Gatty Marine Labo-ratory, University of St. Andrews, Scotland) wereattached to the dorsal pelage of each sea lion usingFive Minute Epoxy (Devcon).

In addition to providing location data, the SRDLswere programmed to maintain a 3-state model of ani-mal activity, determined from time interactions andsurface and depth sensors. Information on behaviorwas processed and compressed into records of behav-ioral states including time spent on shore (i.e. hauledout), extended surface periods, and dive cycles. Sealions were considered hauled out when on shore (dry)for more than 6 min, whereas surface periods weredefined as the time at sea (wet) above a defined divethreshold (<6 m) for more than 6 min. Dives weredefined as being >6 m in depth and >8 s in duration.Due to bandwidth restrictions, dives were sampled atintervals of 4 s, stored in memory, randomly sampled,and transmitted in bouts of 2 or 7 to ensure they were arepresentative sample of the time spent at sea (Fedaket al. 2002). Each day, temperature profiles were alsocollected every 2 h during the deepest dive.

Haulout, surface, and dive records reported by theSRDLs contained the start and end times of unbrokenperiods spent in each state of activity. SRDLs also col-lated summary statistics on the proportion of timespent in each of the 3 states for six 4 h periods d–1

(period 1 = 0:00–3:59 h, period 2 = 4:00–7:59 h, period3 = 8:00–11:59 h, period 4 = 12:00–15:59 h, period 5 =16:00–19:59 h, period 6 = 20:00–23:59 h GreenwichMean Time, GMT). Daily locations and behavioral datafrom SRDLs were obtained through the Service Argossystem (Argos 1996) and decoded in a marine mammalbehavior visualization system (MAMVIS; Fedak et al.1996) by the SMRU. All data were filtered using aswim speed of 2 m s–1 with the algorithm described byMcConnell et al. (1992), and time was used as a meansto interpolate the position of dives along a direct linebetween successive locations.

Standard mapped images of SST (level 3, weeklycomposites, 4.6 km resolution) collected by the moder-ate resolution imaging spectroradiometer (MODIS)instrument aboard NASA’s Aqua satellite wereobtained from the Ocean Color Discipline ProcessingSystem (Campbell et al. 1995). Windows Image Man-ager (6.2; Wimsoft) was used to clip a northern sub-

147

ID Location Latitude Longitude Region Date of Date of last No. of days Sex Age Study(°N) (°E) capture transmission tracked (mo) area (km2)

7820 Buldir 52.340 175.900 WAI 07/03/02 08/13/02 41 M 12 35248.27821 Attu 52.918 172.461 WAI 07/05/02 08/24/02 50 F 12 10834.27822 Attu 52.918 172.461 WAI 07/05/02 07/20/02 15 F 12 14400.88245 Long Island 57.778 –152.416 CGA 02/27/03 03/06/03 7 F 9 –8250 Aiktak 54.183 –164.852 EAI 03/05/03 04/19/03 45 F 21 2869.68252 Aiktak 54.183 –164.852 EAI 03/07/03 03/26/03 19 F 21 2452.29922 Tigalda Rocks 54.139 –164.978 EAI 11/10/03 12/14/03 34 F 5 4130.69923 Tigalda Rocks 54.139 –164.978 EAI 11/10/03 11/22/03 12 F 5 3744.39924 Tigalda Rocks 54.139 –164.978 EAI 11/10/03 01/03/04 54 F 5 1808.49925 Two Headed Isl. 56.897 –153.569 CGA 11/16/03 02/12/04 88 M 5 2869.69926 Two Headed Isl. 56.897 –153.569 CGA 11/16/03 12/26/03 40 F 5 2452.29927 Two Headed Isl. 56.897 –153.569 CGA 11/16/03 01/11/04 56 M 17 4130.69928 Two Headed Isl. 56.897 –153.569 CGA 11/16/03 12/22/03 36 M 5 –9929 Two Headed Isl. 56.897 –153.569 CGA 11/16/03 11/22/03 6 M 5 –10006a Two Headed Isl. 56.897 –153.569 CGA 11/17/03 01/02/04 46 M 5 2154.89930 Long Island 57.781 –152.278 CGA 11/18/03 05/17/04 181 F 29 24891.510007 Kagalaska 51.866 –176.340 CAI 05/06/04 07/07/04 62 M 11 12260.910008 Kagalaska 51.865 –176.340 CAI 05/07/04 07/20/04 74 M 11 5492.610009 Silak Island 51.865 –176.340 CAI 05/07/04 07/23/04 77 M 11 4634.810010 Silak Island 51.865 –176.340 CAI 05/07/04 07/31/04 85 M 11 154125.810011 Silak Island 51.865 –176.340 CAI 05/07/04 08/02/04 87 M 11 9178.510012 Little Tanaga Isl. 51.823 –176.340 CAI 05/16/04 06/22/04 37 F 23 7957.710013 Billingshead 54.290 –165.580 EAI 05/19/04 07/01/04 43 M 11 32665.110014 Akun Island 54.290 –165.580 EAI 05/19/04 06/27/04 39 M 35 44530.7

aThe other 2 study areas for this individual are described in the text (see ‘Results’)

Table 1. Eumetopias jubatus. Capture location, including latitude, longitude (decimal degrees), region, and date of capture, dateof last transmission for satellite transmitters, number of days tracked, sex, age at capture, and study area for juvenile Steller sealions, as indicated by their identification (ID) number. WAI: western Aleutian Islands, CGA: central Gulf of Alaska, EAI: easternAleutian Islands, CAI: central Aleutian Islands, M: male, F: female. Dates given as mo/d/yr. Data for 3 individuals were

discarded because of sensor failure (nos. 8245 and 9928) or limited data (no. 9929)


polar region (including Alaska) from all global remotesensing images, and ENVI (4.0; ITT Visual InformationSolutions) was used to define the datum (i.e. NAD83).ArcInfo was used to convert all remote sensing data toraster grids and project them to an Albers equal-areaconic projection defined for the state of Alaska, USA(ArcGIS 9.0, ESRI). SST data for the week of 28 July to4 August 2002 were not available for analyses.

With the exception of 1 individual (no. 10006), 1study area was devised for each individual by plottingall data collected for the duration of instrument deploy-ment. A minimum convex polygon (MCP) constitutinga simple home range was superimposed on the teleme-try data (Hawths Analysis tools extension, ArcGIS) foreach individual and buffered by 15 km to alleviateedge effects and to account for error of satellite teleme-try positions, which ranges from 0.4 to 17.4 km in areasof Alaska (Fadely et al. 2005). Each buffered MCP wasthen enclosed by a rectangular area that totally cov-ered the geometry of the polygon (Fig. 2). Study areas

were deemed the ‘area of influence’ for each individ-ual assuming these represented individual perception.Weekly SST grids were then clipped to each individualstudy area. The same methods were used to devise 3study areas representing habitat use before, during,and after a trip conducted by sea lion no. 10006, a 5 moold pup that dispersed ~615 km from Two HeadedIsland to Cape St. Elias 11 d after being tagged(Table 1, Fig. 1). The methods were modified for thisindividual because we assumed it followed its motheron that excursion and the entire Gulf of Alaska was notinfluencing its foraging effort.

Data classification. To examine environmental het-erogeneity for this study, weekly categorical mapscomprising SST patches and gradients were derivedfor categorical map analysis. Frontal features, whichare hydrographic features generally defined as aninterface between 2 dissimilar water masses and of-ten characterized by a steep temperature gradient(Etnoyer et al. 2006), were defined as cells where the

148

Fig. 1. Eumetopias jubatus. Geographical regions in Alaska comprising the western distinct population segment (wDPS) of Stellersea lions. Asterisks (*) indicate regions where sea lions (n = 24) were captured. Regions include the western Aleutian Islands(WAI), central Aleutian Islands (CAI), Eastern Aleutian Islands (EAI), western Gulf of Alaska (WGA), central Gulf of Alaska

(CGA), and eastern Gulf of Alaska (EGA)


SST gradient was greater or less than1 SD from the mean gradient of the studyarea for each individual (Moore et al.2002); hence, this classification schemeresulted in 3 data classes (Fig. 3). To cre-ate gradients, defined as a change in aproperty across a defined spatial extentrepresenting a pattern of continuous vari-ation of a single focal variable (Lovettet al. 2003, White & Brown 2003), theslope function (Spatial Analyst extension,ArcGIS 9.1) was used to calculate the rate(degrees) of maximum change in SSTfrom each data cell and its 8 neighbors foreach weekly grid per individual.

After data classification, FRAGSTATS3.3 (McGarigal & Marks 1995) was usedto determine the number of patches cor-responding to each data class, where apatch was defined as a contiguous groupof cells of the same mapped categoryusing an 8-neighbor rule (i.e. 2 grid cellsof the same cover type are consideredpart of the same patch if they are adjacentor diagonal neighbors; Forman & Godron1986, Turner et al. 2001). Three landscapemetrics, which measure the aggregateproperties of the entire grid mosaic, wereused to characterize heterogeneity of SSTfor each remote sensing week, includingpatch density (PD), Simpson’s diversityindex (SIDI), and area-weighted mean

149

Fig. 2. Eumetopias jubatus. An example illustrating how study areas weredevised for 21 juvenile Steller sea lions. A minimum convex polygon wassuperimposed on pooled, filtered telemetry data for each individual,buffered by 15 km, and then enclosed by a rectangular area that totally cov-ered the geometry of the polygon. Weekly sea surface temperature grids

were then clipped to each individual study area

Fig. 3. Eumetopias jubatus. The methodology used for classifying (A) weekly sea surface temperature (SST) data entailed (B)deriving SST gradients by calculating the rate of maximum change in SST from each data cell and its 8 neighbors. (C) Frontalfeatures were then defined as cells where the SST gradient was greater (class 3) or less than (class 1) 1 SD from the mean gradient

of the study area for each individual


fractal dimension (AMFD). We chose to examine PDand SIDI of SST because Lander et al. (2009) found thatthese covariates may be linked to regional populationtrends of Steller sea lions, whereas AMFD provides ameasure of complexity.

PD (total number of patches area–1) was computedfor individual study areas and reported as number ofpatches per 100 ha. The AMFD was calculated usingthe following metric:

(1)

where n = total number of patches, pi = perimeter (m)of patch i, and ai = area (m2) of patch i. This metricincreases as shape complexity of patches within thelandscape mosaic increases and is advantageousbecause it represents complexity (or departure fromEuclidean geometry) across a range of spatial scales(i.e. patches). This metric was weighted because smallmaps are more prone to effects caused by map borders;a greater proportion of patches are truncated at theedges of the map (Hargis et al. 1997). Additionally,SIDI was calculated for each week using the followingmetric:

(2)

where Pi = the proportion of the landscape occupied byclass type i, and c = the number of classes present. SIDIrepresents the probability that any 2 pixels selected atrandom would be different patch types (McGarigal etal. 2002), and SIDI = 0 when the area is dominated by1 patch (no diversity) and approaches 1 as the numberof different patch types increases and the proportionaldistribution of area among patch types becomes moreeven. Relative to other diversity indices (e.g. Shan-non’s diversity index), SIDI is less sensitive to the pres-ence of rare patch/group types, so more weight isplaced on common patch/group types. This character-istic coupled with a fairly consistent number of mapclasses (i.e. richness) across regions enabled us toavoid problems associated with having study areas ofdifferent sizes.

Four class metrics were computed only for patchescorresponding to map classes representing frontalfeatures to further elucidate how these specificpatches influenced the foraging effort of sea lions. Inaddition to PD and AMFD, area-weighted mean classarea (AMCA) was calculated using the followingmetric:

(3)

where aij = area (m2) of patch ij and n = number ofpatches corresponding to frontal features. The metricis divided by 10 000 to convert to ha. Additionally, thepercentage of landscape (PLAND) comprising theseclasses was also calculated (Σ area of each designatedclass patch/area × 100). This estimate approaches 100as the entire image becomes composed of a singlepatch (McGarigal et al. 2002).

Statistical analysis. Trip durations for individual sealions were calculated using the departure and arrivaltimes for trips at sea provided in the haulout records.Mean trip duration and percentage of time spent div-ing for each sea lion were calculated for each remotesensing week. Trips that straddled 2 weeks wereassigned to the week containing the greater proportionof the trip. Maximum trip duration wk–1 was alsoexamined for each individual. All partial weeks wereused in analyses.

Mean and maximum trip duration wk–1 were log-transformed after conducting Kolmogorov-Smirnovtests for normality (SPSS 13.0) and examining histo-grams and q-q plots of the response variables (R 2.4.1,R Foundation for Statistical Computing). Generalizedlinear models (GLMs) with a Gaussian error, identitylink, and first-order autoregressive correlation struc-ture were used to examine mean and maximum tripduration wk–1, and percentage of time spent divingwk–1 with respect to the covariates region, age (atcapture), and environmental metrics described above(geepack 1.0-10, R 2.4.1; Liang & Zeger 1986, Zeger &Liang 1986). Individual sea lion was used as thegrouping variable. Regions were coded as individualindicator variables, and the central Aleutian Islands(CAI) region was used as the control. These modelsaccounted for longitudinal data, repeated measures,and correlated responses within each sea lion for theresponse variables, and possible time-dependentcovariates. This approach is advantageous because itaccounts for serial correlation in the response, isrobust to deviations from normality, and employs aquasi-likelihood approach to provide generalized esti-mating equations (GEE; Yan & Fine 2004, Halekoh etal. 2006).

Sex was not examined as a predictor variablebecause preliminary data analyses indicated that theresponse variables did not differ between males andfemales. Year or season also were not examined dueto paucity of data and because previous studies indi-cated that year was not an important factor for pre-dicting trip duration for Steller sea lions (Call et al.2007). Optimal models were developed by using abackward stepwise elimination procedure to removenon-significant terms one at a time based on the sig-nificance levels of Wald test statistics (p ≤ 0.15 formodel retention, p ≤ 0.05 for significance). Separate

AMCA =

⎛

⎝

⎜⎜⎜

⎞

⎠

⎟⎟⎟

⎡

⎣

⎢⎢⎢⎢

⎤

⎦

⎥⎥⎥⎥=

= ∑∑ a

a

aij

ij

ijj

n

j

n

11

100 000

SIDI = −=

∑1 2

1

Pii

c

AMFD = ⎡⎣⎢

⎤⎦⎥

⎛

⎝

⎜⎜⎜

⎞

⎠

⎟⎟⎟

⎧

⎨=

∑2 0 25

1

ln( . )ln

pa

a

ai

i

i

ii

n⎪⎪

⎩⎪

⎫

⎬⎪

⎭⎪=

∑i

n

1

150


analyses were conducted for each response variablefor the 2 different groups of metrics (i.e. landscapeand class metrics; n = 6 models). Lastly, to assess theappropriateness of final models, q-q plots and resid-ual plots were examined (R 2.4.1).

RESULTS

From 3 July 2002 to 19 May 2004, 24 juvenile Stellersea lions (5 to 35 mo old) were captured at rookeriesor haulout sites within the western (n = 3), central (n =6), and eastern (n = 7) Aleutian Islands, and the cen-tral Gulf of Alaska (n = 8; Fig. 1, Table 1). Data for 3individuals were discarded because of sensor failure(nos. 8245 and 9928) or limited data (no. 9929). Instru-ment deployment for the remaining 21 sea lions lastedfrom 12 to 181 d. Study areas for individual sea lionsranged from 1808.4 to 154 125.8 km2 (Table 1). The3 study areas representing habitat use before, dur-ing, and after the trip conducted by sea lion no. 10006were 2154.8 km2, 216 828.9 km2, and 5039.4 km2,respectively.

Trip durations

Overall, the duration of trips (n = 1475) for all indi-viduals (n = 21) ranged from 0.1 to 177.1 h (mean ± SD= 8.6 ± 14.8 h). Average trip duration ind.–1 rangedfrom 3.0 to 28.6 h, whereas maximum trip durationranged from 14.1 to 177.1 h (68.4 ± 47.8 h; Table 2).Mean trip duration wk–1 ranged from 0.9 to 117.6 h(12.3 ± 14.5 h, n = 151 wk, range = 2–23 wk ind.–1).

Activity patterns

Summary statistics for the 3 behavioral states wereobtained for 896 complete records (full 24 h periods;range = 7–124 records ind.–1, n = 21 ind.). On a weeklybasis, sea lions spent an average ± SD of 55.7 ± 18.3%(range = 0.0–98.3%) of their time on shore, whereas44.3 ± 18.3% was spent at sea (range = 1.7–100.0%;n = 145 wk, range = 2–23 wk ind.–1). While at sea, sealions spent an average of 11.2 ± 6.8% (range =0.0–28.4%) of their time diving and 33.1 ± 14.2%(range = 1.5–83.1%) of their time at the surface. How-ever, these activities varied considerably within andamong individuals (Table 3).

Time of day during which most diving activityoccurred varied among individuals, but on average,sea lions spent a greater proportion of time diving atnight during period 3 (Table 4), followed by crepuscu-lar periods (periods 2 and 4; Table 4). A greater propor-

tion of time spent diving consistently occurred at nightfor sea lions captured in the CAI and eastern AleutianIslands (EAI) during May 2004.

From a regional perspective, sea lions from the west-ern Aleutian Islands (WAI) spent an average ± SD of47.1 ± 17.9% of the day on shore, 7.3 ± 3.4% diving,and 45.6 ± 16.9% at the surface (traveling, resting, ordiving to depths <6 m). Sea lions from the CAI spent anaverage of 57.0 ± 13.0% of the day on shore, 13.0 ±5.4% diving, and 30.0 ± 9.0% at the surface. Sea lionsfrom the EAI spent an average of 64.4 ± 16.0% of theday on shore, 8.3 ± 5.7% diving, and 27.3 ± 13.0% atthe surface. Sea lions from the central Gulf of Alaska(CGA) spent an average of 49.9 ± 22.2% of the day onshore, 12.3 ± 8.4% diving, and 37.8 ± 15.9% at thesurface.

GLMs

Predictors of weekly mean and maximum trip dura-tions, and percentage of time spent diving variedamong models (Table 5). Foraging effort as indicatedby all response variables increased with age for allmodels (Table 5). Mean trip duration did not differ sig-nificantly among regions, whereas maximum tripduration was greatest in the WAI and differed signifi-cantly from the CAI (Table 5). Additionally, sea lionsfrom the WAI and the EAI spent less time diving wk–1

151

ID No. of Trip duration (h)trips Mean SD Min. Max.

7820 51 8.4 15.5 0.1 105.87821 52 14.1 13.2 0.1 48.37822 29 4.2 10.2 0.1 51.08250 81 4.6 4.5 0.1 23.68252 28 4.6 3.8 0.1 14.19922 73 3.0 6.6 0.1 50.69923 14 11.2 25.5 0.1 98.29924 71 5.7 8.0 0.1 53.79925 90 3.4 3.9 0.1 20.89926 48 5.3 5.1 0.2 29.09927 29 28.6 42.5 0.1 177.110006 90 3.7 11.5 0.1 109.79930 145 19.4 25.3 0.1 151.710007 84 6.0 6.2 0.1 31.510008 108 7.5 6.5 0.1 36.510009 95 8.7 7.4 0.1 46.510010 123 8.1 17.7 0.1 156.310011 87 13.0 9.7 0.1 47.310012 62 6.0 6.0 0.1 35.010013 60 8.5 11.0 0.1 54.310014 55 8.4 14.8 0.1 94.6

Table 2. Eumetopias jubatus. Steller sea lion identification(ID), sample size (number of trips at sea), and mean ± SDand range (minimum and maximum) of trip duration (h) for

individual deployment periods


than did sea lions from the CAI and presumably theCGA (Table 5). Percentage of time spent diving wk–1

increased with an increase in fractal dimension at boththe landscape (mean ± SD AMFD = 1.097 ± 0.036,range = 1.026–1.182) and class (i.e. localized frontal

features) scales (mean ± SD, AMFD = 1.059 ± 0.024,range = 1.010–1.128; Table 5). Weekly mean trip dura-tions were inversely related to PD of frontal features(mean ± SD PD = 0.0035 ± 0.003 patches 100 ha–1,range = 0.0002–0.0096; Table 5).

152

ID No. of days Diving Range At surface Range Hauled out Range

7820 23 6.9 ± 6.9 0.0–28.8 35.9 ± 23.5 0.3–89.5 57.2 ± 28.5 0.0–99.77821 34 7.8 ± 4.2 0.0–17.7 51.8 ± 26.1 6.1–99.2 40.4 ± 27.6 0.0–93.97822 9 4.7 ± 3.8 0.0–8.9 36.7 ± 32.7 2.4–91.3 58.6 ± 35.8 0.0–97.68250 39 9.7 ± 6.8 0.0–23.3 24.8 ± 15.0 0.0–72.6 65.6 ± 19.6 11.1–100.08252 15 8.2 ± 7.6 0.0–22.7 20.9 ± 17.9 0.0–63.6 71.0 ± 23.5 27.8–100.09922 31 3.9 ± 4.2 0.0–16.4 25.1 ± 20.9 0.0–99.8 70.9 ± 23.0 0.0–100.09923 7 2.3 ± 2.7 0.0–5.8 35.5 ± 31.2 0.0–99.8 62.2 ± 30.9 0.0–100.09924 49 3.4 ± 3.7 0.0–18.0 21.4 ± 18.4 0.0–61.4 75.3 ± 20.8 20.6–100.09925 33 4.1 ± 5.0 0.0–17.2 24.2 ± 14.7 0.0–54.8 71.7 ± 17.8 34.6–100.09926 36 3.0 ± 4.1 0.0–16.6 23.6 ± 18.8 0.0–78.1 73.5 ± 20.9 5.3–100.09927 48 18.7 ± 12.3 0.0–39.4 41.6 ± 28.1 0.0–88.9 39.7 ± 38.3 0.0–100.010006 41 1.3 ± 2.5 0.0–11.3 29.9 ± 25.3 0.0–99.0 68.8 ± 26.6 0.0–100.09930 124 17.5 ± 7.9 0.0–39.9 44.9 ± 23.8 0.0–86.0 37.5 ± 28.8 0.0–100.010007 56 8.6 ± 7.7 0.0–26.4 21.3 ± 12.8 0.0–69.1 70.1 ± 17.7 29.9–100.010008 62 12.0 ± 5.5 0.0–31.4 31.4 ± 13.5 0.0–59.3 56.6 ± 17.4 12.2–100.010009 68 12.4 ± 6.2 0.0–27.7 32.4 ± 16.0 0.0–75.2 55.2 ± 21.0 0.0–100.010010 61 11.5 ± 7.7 0.0–35.9 31.4 ± 21.0 0.0–86.6 57.1 ± 26.5 0.0–100.010011 64 19.1 ± 6.7 0.0–35.9 31.6 ± 14.4 0.0–63.8 49.3 ± 19.4 6.3–100.010012 31 12.6 ± 3.0 7.7–22.7 24.9 ± 12.6 9.8–64.6 62.4 ± 13.9 25.1–80.910013 34 14.3 ± 7.2 0.0–29.3 28.8 ± 19.0 1.6–77.8 56.9 ± 23.9 0.0–98.410014 31 11.5 ± 5.6 2.6–28.8 32.6 ± 21.3 6.4–83.1 55.9 ± 25.5 0.0–91.0

Table 3. Eumetopias jubatus. Steller sea lion identification (ID), number of days of data (number of records), and summary statis-tics for activity patterns, including mean ± SD percentage of time spent diving, at the surface, and hauled out on shore for the

duration of the deployment period

ID 1 2 3 4 5 6(00:00–3:59 h) (04:00–7:59 h) (08:00–11:59 h) (12:00–15:59 h) (16:00–19:59 h) (20:00–23:59 h)

7820 3.4 2.6 15.2 10.9 4.2 5.37821 9.9 5.4 1.4 1.6 13.2 15.17822 5.5 1.6 7.7 9.0 1.4 3.08250 0.2 8.1 20.1 25.9 2.5 1.28252 0.5 6.9 9.9 16.0 12.5 3.29922 1.2 6.9 3.8 6.4 3.6 1.89923 0.4 7.9 2.7 2.2 0.7 0.09924 2.0 5.4 4.1 3.9 2.8 1.99925 2.7 1.3 1.8 4.0 6.9 4.09926 4.0 2.0 2.7 4.1 1.7 3.29927 10.8 24.5 30.2 30.1 13.0 3.410006 1.9 1.4 0.9 1.3 1.5 1.09930 9.1 28.3 33.9 23.1 5.6 5.310007 0.6 2.2 23.2 19.0 5.9 1.010008 0.9 2.6 34.7 26.6 4.1 3.110009 2.0 4.7 29.5 25.4 7.9 4.810010 4.8 6.5 27.5 17.5 6.5 6.510011 4.7 5.2 44.6 38.4 14.6 7.010012 0.0 4.0 48.5 21.8 0.9 0.510013 3.4 15.3 40.8 12.2 6.4 7.910014 4.9 7.6 37.5 10.4 4.6 3.9

Mean ± SD 3.5 ± 3.2 7.2 ± 7.2 20.0 ± 16.2 14.8 ± 10.8 5.7 ± 4.3 4.0 ± 3.4

Table 4. Eumetopias jubatus. Mean percentage of time spent diving within 6 periods of the day for 21 juvenile Steller sea lions, asindicated by their identification number (ID). Time of day refers to Greenwich Mean Time (GMT). Overall mean ± SD are

provided for each period


DISCUSSION

Similar to other studies of otariids (McCafferty et al.1998, Baker & Donohue 2000, Baylis et al. 2005),including Steller sea lions (Raum-Suryan et al. 2004,Fadely et al. 2005, Pitcher et al. 2005, Call et al. 2007,Rehberg & Burns 2008), age was a significant factor forpredicting the amount of time spent at sea and diving.These results were not surprising, as an increase inexperience coupled with the development of physio-logical abilities (muscle and blood oxygen stores andthermal tolerance) occur with age and growth (Rich-

mond et al. 2005, 2006), enabling sealions to remain at sea for longer peri-ods of time. Summary statistics ofactivity patterns were also similar tofindings of other researchers whofound that on average, Steller sea lionsspent 44% of their time at sea(Rehberg 2005, Call et al. 2007). Ourresults also corroborated other studies(Loughlin et al. 2003, Fadely et al.2005, Call et al. 2007) indicating thatmost diving activity of pups and juve-niles occurred during the night whensea lions may have been foraging onshallow vertically migrating prey suchas walleye pollock Theragra chalco-gramma. Pollock tend to school atdepth during daytime hours and dis-perse as they rise in the water columnat night (Sinclair et al. 1994). Theseresults imply that sea lions spent theremainder of time at sea conductingother activities (unless foraging acti-vities occurred within 6 m of thesurface).

Sea lions from the WAI spent lesstime diving while at sea, yet maximumtrip durations were greatest for thisgroup of animals. These results cou-pled with the proportion of time spentat the surface possibly indicate thatmore time was allocated to travelingand less to resource utilization. Rela-tive to the other 2 regions, sea lionsfrom the EAI also spent less timediving during the week, but their div-ing behavior constituted a greater pro-portion of their overall time at sea(Table 3). Assuming that foragingindices are indicators of variability inmarine resources (Boyd 1999) andreflect differences in the cost of preyacquisition (Costa et al. 1989, Merrick

& Loughlin 1997), these data suggest that sea lionsfrom the EAI were either more efficient foragers thanthose from the WAI or resources were more attainablein the EAI than the WAI. Results of Lander et al. (2009),which indicated that diet diversity and habitat diver-sity were greater for the EAI than the WAI, support thelatter notion if in fact diet and habitat diversity reflectprey diversity and abundance in the environment.Results should be interpreted with caution, however,because regional differences in diving effort may havebeen hampered by foraging behavior, age, season, orinteraction effects (which were not examined due to

153

Response variable Coefficient SE Wald p Model pVariables retained estimate

LandscapeMean trip duration wk–1

Intercept* 2.510 0.112 498.484 0.000 4.172 × 10–8

Age* 0.004 0.002 6.378 0.012Region (EAI) –0.214 0.121 3.136 0.077

Max. trip duration wk–1

Intercept 2.860 0.091 996.593 0.000 4.649 × 10–10

Age* 0.004 0.002 5.465 0.019Region (WAI)* 0.289 0.055 27.619 1.477 × 10–7

% time diving wk–1

Intercept* –32.569 19.945 2.666 0.001 1.547 × 10–7

Age* 0.098 0.021 21.465 3.603 × 10–6

Region (EAI)* –4.194 2.157 3.782 0.052Region (WAI)* –4.223 1.648 6.568 0.010SIDI –6.549 3.936 2.769 0.096AMFD* 37.819 18.315 4.264 0.039

ClassMean trip duration wk–1

Intercept* 2.481 0.163 231.250 0.000 1.468 × 10–7

Age* 0.004 0.002 6.067 0.014Region (EAI) –0.209 0.119 3.083 0.079PD* –33.014 16.229 4.138 0.042PLAND 0.006 0.004 2.984 0.084

Max. trip duration wk–1

Intercept* 3.028 0.114 704.327 0.000 2.721 × 10–8

Age* 0.004 0.001 8.934 0.003Region (EAI) –0.161 0.102 2.491 0.114Region (WAI)* 0.196 0.055 12.920 0.000PD –21.547 11.220 3.688 0.055

% time diving wk–1

Intercept –24.666 16.752 2.168 0.141 4.864 × 10–7

Age* 0.110 0.021 28.274 1.053 × 10–7

Region (CGA) –3.486 2.303 2.292 0.130Region (EAI)* –5.712 2.090 7.472 0.006Region (WAI)* –5.510 1.763 9.768 0.002AMFD* 30.284 15.732 3.706 0.054

Table 5. Eumetopias jubatus. Results of 6 stepwise generalized estimating equa-tions used to examine 3 predictor variables (weekly, mean and maximum tripduration, and percentage of time spent diving by 21 juvenile Steller sea lions)with respect to age, region, and metrics (landscape and class) of sea surfacetemperature heterogeneity. All models were significant. *p ≤ 0.05. EAI: easternAleutian Islands, WAI: western Aleutian Islands, SIDI: Simpson’s diversity in-dex, AMFD: area-weighted mean fractal dimension, PD: patch density, PLAND:

percentage of landscape, CGA: central Gulf of Alaska


lack of degrees of freedom). Various foraging tacticsare used by Steller sea lions to feed on different typesof prey as a result of differences in prey characteristicssuch as size, age, and behavior. Because dominantprey types vary among regions (Sinclair & Zeppelin2002, Lander et al. 2009), it is possible that foragingeffort among regions was influenced by differentstrategies used for the pursuit, capture, and consump-tion of disparate prey. Furthermore, sample sizes werelimited and all age classes of sea lions were not equallyrepresented among regions. The weaning status of sealions captured for this study was unknown, and indi-viduals that were still nursing may have influenced ourinterpretation of the results if they were less motivatedto find food resources affected by environmental het-erogeneity (McIntyre & Wiens 1999). Unfortunately,lack of long-term data for most animals prevented usfrom examining how these behaviors changed overlonger periods of time.

Other studies have indicated that the amount andspatial arrangement of resources and habitat constrainthe location, movement, and foraging dynamics ofother species (Crist et al. 1992, Ferguson et al. 1998,With et al. 1999). Our results indicated that divingactivity of Steller sea lions increased as the shape com-plexity of localized frontal features within the studyareas increased, whereas average weekly trip durationwas inversely related to PD of those features. Assum-ing that increases in foraging effort and trip durationsare predicted if prey availability is reduced (Trillmich& Ono 1991, Lunn et al. 1993, Hood & Ono 1997, Mer-rick & Loughlin 1997, McCafferty et al. 1998, Georgeset al. 2000, Melin 2002, Weimerskirch et al. 2003), thiscombination of results suggests that sea lions had amore difficult time obtaining resources as the structureof SST frontal features became more complex, but mayhave attained resources more quickly as the number ofpatches comprising frontal features increased. Thus,aggregated, dense frontal features were probably con-ducive to foraging effort as opposed to time periodswhen these features were forming or dissipating. Thelack of significance for AMCA and PLAND furthersuggests that complexity, rather than size, of frontalfeatures influenced diving behavior.

Frontal zones, including thermal fronts, tend to beregions of enhanced primary productivity relative tosurrounding areas (Graham et al. 2001, Okkonen et al.2003, Bradshaw et al. 2004) and are important to othermarine mammals (Hindell et al. 1991, Sinclair et al.1994, Moore et al. 2002, Etnoyer et al. 2006, Doniol-Valcroze et al. 2007), seabirds (Hunt et al. 1999), seaturtles (Etnoyer et al. 2006), pelagic fishes (Royer et al.2004), and other marine fauna (Graham et al. 2001).These predators likely concentrate at fronts due to foodavailability or thermal constraints, which may be evo-

lutionary, ecological, or physiological (Brandt 1993).For example, metabolic rates and gut passage rates ofsome fish species are affected by water temperature(Gillooly et al. 2001). It is also believed that water tem-perature influences the availability, behavior, spawn-ing, and survival of forage fish (Bailey et al. 1995),which are important prey species of Steller sea lions.Additionally, frontal features may be beneficial be-cause they have more thermal habitats per unit areathan surrounding waters and have a greater probabil-ity of encompassing a preferred range of temperature(Brandt et al. 1980), thereby meeting energetic de-mands and other requirements needed for survival.

Unlike diving activity, trip duration did not increaseas fractal dimension of frontal features increased,despite the idea that the 2 variables generally coincideor that it may take longer to navigate around irregularpatches. However, assuming frontal features at thescales examined were used by Steller sea lions, otherfactors such as permeability of features, proximityamong features, or location of features relative tohaulouts or rookeries may have affected the results.For example, the spatial arrangement and shape offeatures presumably reflect their connectivity, but notnecessarily their boundary characteristics. Althoughdiving activity appeared to be more efficient duringtimes when study areas contained connected, aggre-gated features, the configuration of features may nothave affected trip durations if the features themselvesdid not pose a barrier to sea lions. Additionally, if sealions tend to target frontal features of a specific config-uration, which may be indicative of the strength of anunderlying process or the concentration of a prey field,then trip durations may not have been related to mea-sures of AMFD or the amount of diving activitybecause those features can potentially occur anywherewithin a given study area mosaic.

In contrast to results reported by Lander et al. (2009),which indicated that regional patterns of SST diversitywere fairly consistent with regional population trajec-tories of Steller sea lions, indices of foraging effortexamined for this study were not related to SST diver-sity. Although this was unexpected, our results andthose of Lander et al. (2009) appear to conform to thetheory that landscape composition (e.g. habitat diver-sity) has large, direct effects on population dynamicsand persistence (possibly through direct effects onreproduction and mortality), whereas landscape con-figuration (e.g. fractal dimension) affects populationdynamics indirectly through its effects on among-patch movements (Fahrig & Nuttle 2003). Although aninteresting observation, additional empirical evidenceis needed to assess this hypothesis, which differsslightly from classical metapopulation theory, anapproach that predicts that survival of endangered

154


populations is dominated by stochastic events and thespatial arrangement of suitable habitat (Hanski 1998).

The National Marine Fisheries Service (NMFS)Recovery Plan for Steller sea lions (NMFS 2008) sug-gests that critical habitat should be enhanced to incor-porate the spatial and temporal variation of essentialoceanographic features that potentially influence thedistribution and abundance of prey and ultimate con-servation of Steller sea lions. It is therefore necessaryto determine which features constitute essential habi-tat, how those features persist over time, and how sealions (and their prey) exploit those features. For thisstudy, we described how the foraging effort of Stellersea lions was related to the heterogeneity of small-scale surface temperature gradients. However, wehave an incomplete understanding of how exactlythose features were truly perceived or used by sealions because the distributions of sea lion locations (ordives) were not analyzed with respect to the exactlocality of defined features due to the spatial resolutionof the telemetry and remote sensing data. Further-more, the environmental patterns demarcated for thisexercise may have been sensitive to changes in scaleand decision rules used to classify the data (e.g. patchdefinitions and parameter inputs; Turner et al. 1989).Similar studies in the future will undoubtedly benefitfrom simultaneously sampling the prey environment,the sophistication of ocean observing satellites, and theadvancement of GPS technology.

Linking underlying processes to observed environ-mental patterns is essential for understanding the func-tional relevance of our results. Tidal advection, weatherconditions, wind-forcing, and bottom topography canall contribute to the formation of frontal features andcreate patterned heterogeneity in the marine environ-ment. At larger meso-scales, these gradients and/orfronts can represent the boundaries between differentwater masses or they may be indicative of other meso-scale features such as transient eddies (Ladd et al.2005). It has been suggested that northern fur seals Cal-lorhinus ursinus, and possibly Steller sea lions onlonger pelagic excursions, are attracted to these fea-tures (Sinclair et al. 1994, Fadely et al. 2005, Ream et al.2005), which likely concentrate productivity and prey,facilitate movement, and increase foraging opportuni-ties (Ream et al. 2005). At smaller spatial scales, such asheadlands and islands where Steller sea lions tendto congregate more often, currents may interact withtopography to produce complex 3-dimensional sec-ondary flows that result in physical and biologicalfronts that can influence the distribution of many or-ganisms (Wolanski & Hamner 1988). Other 2-dimen-sional horizontal features such as coastal fronts are alsoassociated with vertical motion and heterogeneity (Ab-bott 1993) and may be surface expressions of greater

subsurface gradients (Roughan et al. 2005). Althoughthere were cases when plotted temperature profileswithin or near the border of classified features sup-ported this idea, the majority of temperature profilescollected during this study indicated that sea lions wereforaging within the mixed layer. Hence, small-scalesurface gradients, which typically result from surfacewind stress (Langmuir circulation) or internal tides andare often marked by an entrainment of surface debris,buoyant particulates, and plankton (Wolanski & Ham-ner 1988, The Open University 2001), should be investi-gated further. Thin layers, which occur in coastal areas,contain high concentrations of living organisms, andpossibly result from similar processes (Franks 1995,Johnston et al. 2009), also warrant future attention.

Typing small-scale features used by juvenile Stellersea lions inevitably will entail the use of additionalsensors to collect precise in situ measurements ofoceanographic variables in coastal waters unavailableto satellite remote sensing platforms. Recent advancesin biologging instruments, including the developmentof a conductivity-temperature-depth SRDL (CTD-SRDL), are promising and have allowed examinationof the behaviors of larger marine mammals in the con-text of a 3-dimensional environment (Biuw et al. 2007),mapping major fronts (Charrassin et al. 2008), and ana-lyzing data in innovative ways (Weise et al. in press).

Acknowledgements. We thank the entire staff of the AlaskaEcosystems Program at the National Marine Mammal Labora-tory, Alaska Fisheries Science Center, NOAA Fisheries, theAlaska Department of Fish and Game, and the crews of theMV ‘Pacific Star,’ MV ‘Tiglax,’ and MV ‘Woldstad’ for all oftheir support and assistance in capturing and instrumentingsea lions. J. Benson, L. Delwiche, A. Greig, D. Johnson, A.Zerbini, the University of Washington (UW) Biostatistics con-sulting group, and the University of St. Andrew’s Sea Mam-mal Research Unit also provided valuable assistance withdata processing and analysis. This manuscript was greatlyimproved through reviews by J. Ver Hoeuf, W. Testa, G.Duker, J. Lee, and 3 anonymous reviewers. This work wasconducted under Federal Marine Mammal Permit no. 782-1532 and UW IACUC protocol no. 2887-09. The use of trade,product, or firm names in this publication is for descriptivepurposes only and does not imply endorsement by the USGovernment.

LITERATURE CITED

Abbott MR (1993) Phytoplankton patchiness: ecologicalimplications and observation methods. In: Levin SA, Pow-ell TM, Steele JH (eds) Patch dynamics. Springer-Verlag,New York, p 37–49

Argos (1996) User’s manual, satellite based data collectionand location system. Service Argos, Landover, MD

Atkinson S, DeMaster DP, Calkins DG (2008) Anthropogeniccauses of the western Steller sea lion Eumetopias jubatuspopulation decline and their threat to recovery. MammalRev 38:1–18

155


Bailey KM, Macklin SA, Reed RK, Brodeur RD and others(1995) ENSO events in the northern Gulf of Alaska, andeffects on selected marine fisheries. Calif Coop OceanFish Invest Rep 36:78–96

Baker JD, Donohue MJ (2000) Ontogeny of swimming anddiving in northern fur seal (Callorhinus ursinus) pups. CanJ Zool 78:100–109

Baylis AMM, Page B, Peters K, McIntosh R, Mckenzie J,Goldsworthy S (2005) The ontogeny of diving behaviour inNew Zealand fur seal pups (Arctocephalus forsteri). Can JZool 83:1149–1161

Benson AJ, Trites AW (2002) Ecological effects of regimeshifts in the Bering Sea and eastern North Pacific Ocean.Fish Fish 3:95–113

Biuw M, Boehme L, Guinet C, Hindell M and others (2007)Variations in behavior and condition of a Southern Oceantop predator in relation to in situ oceanographic condi-tions. Proc Natl Acad Sci USA 104:13705–13710

Boyd IL (1999) Foraging and provisioning in Antarctic furseals: interannual variability in time-energy budgets.Behav Ecol 10:198–208

Boyd IL, Hawker EJ, Brandon MA, Staniland IJ (2001) Mea-surement of ocean temperatures using instruments carriedby Antarctic fur seals. J Mar Syst 27:277–288

Bradshaw CJA, Higgins J, Michael KJ, Wotherspoon SJ, Hin-dell MA (2004) At-sea distribution of female elephantseals relative to variation in ocean surface properties.ICES J Mar Sci 61:1014–1027

Braham HW, Everitt RD, Rugh DJ (1980) Northern sea lionpopulation decline in the eastern Aleutian Islands. J WildlManag 44:25–33

Brandt SB (1993) The effect of thermal fronts on fish growth: abioenergetics evaluation of food and temperature. Estuar-ies 16:142–159

Brandt SB, Magnuson JJ, Crowder LB (1980) Thermal habitatpartitioning of fishes in Lake Michigan. Can J Fish AquatSci 37:1557–1564

Cadenasso ML, Pickett STA, Weathers KC, Jones CG (2003) Aframework for a theory of ecological boundaries. Bio-science 53:750–758

Call KA, Loughlin TR (2005) An ecological classification ofAlaskan Steller sea lion (Eumetopias jubatus) rookeries:a tool for conservation/management. Fish Oceanogr 14(Suppl 1):212–222

Call KA, Fadely BS, Greig A, Rehberg MJ (2007) At-sea andon-shore cycles of juvenile Steller sea lions (Eumetopiasjubatus) derived from satellite dive recorders: a compari-son between declining and increasing populations. Deep-Sea Res II 54:298–310

Campbell JW, Blaisdell JM, Darzi M (1995) Level-3 SeaWiFSdata products: spatial and temporal binning algorithms.In: Hooker SB, Firestone ER, Acker JG (eds) NASA TechMemo 104566, Vol 32. NASA Goddard Space FlightCenter, Greenbelt, MD

Charrassin JB, Hindell M, Rintoul SR, Roquet F and others(2008) Southern Ocean frontal structure and sea-ice for-mation rates revealed by elephant seals. Proc Natl AcadSci USA 105:11634–11639

Costa DP, Croxall JP, Duck C (1989) Foraging energetics ofAntarctic fur seals in relation to changes in prey availabil-ity. Ecology 70:596–606

Crist TO, Guertin DS, Wiens JA, Milne BT (1992) Animal move-ment in heterogeneous landscapes: an experiment withEleodes beetles in shortgrass prairie. Funct Ecol 6:536–544

DeMaster D, Atkinson S (2002) Steller sea lion decline: Is itfood? II. University of Alaska Sea Grant, AK-SG-0202,Fairbanks, AK

Doniol-Valcroze T, Berteaux D, Larouche P, Sears R (2007)Influence of thermal fronts on habitat selection by fourrorqual whale species in the Gulf of St. Lawrence. MarEcol Prog Ser 335:207–216

Etnoyer P, Canny D, Mate BR, Morgan LE, Ortega-Ortiz JG,Nichols WJ (2006) Sea-surface temperature gradientsacross blue whale and sea turtle foraging trajectories offthe Baja California Peninsula, Mexico. Deep-Sea Res II 53:340–358

Fadely BS, Robson BW, Sterling JT, Greig A, Call KA (2005)Immature Steller sea lion (Eumetopias jubatus) dive activ-ity in relation to habitat features of the eastern AleutianIslands. Fish Oceanogr 14:243–258

Fahrig L, Nuttle WK (2003) Population ecology in spatiallyheterogeneous environments. In: Lovett GM, Jones CG,Turner MG, Weathers KC (eds) Ecosystem function inheterogeneous landscapes. Springer Science + BusinessMedia, New York, p 95–118

Fedak MA, Lovell P, McConnell BJ (1996) MAMVIS: Amarine mammal behaviour visualization system. J VisualComput Animation 7:141–147

Fedak MA, Lovell P, McConnell B, Hunter C (2002) Overcom-ing the constraints of long range radio telemetry from ani-mals: getting more useful data from smaller packages.Integr Comp Biol 42:3–10

Ferguson SH, Taylor MK, Born EW, Messier F (1998) Fractals,sea-ice landscape and spatial patterns of polar bears.J Biogeogr 25:1081–1092

Forman RT, Godron M (1986) Landscape ecology. Wiley Pub-lishers, New York

Franks PJS (1995) Thin layers of phytoplankton: a model offormation by near-inertial wave shear. Deep-Sea Res I 42:75–91

Fritz LW, Hinckley S (2005) A critical review of the regimeshift — ‘junk food’ — nutritional stress hypothesis for thedecline of the western stock of Steller sea lion. Mar MammSci 21:476–518

Fritz L, Lynn M, Kunisch E, Sweeney K (2008) Aerial, ship,and land-based surveys of Steller sea lions (Eumetopiasjubatus) in Alaska, June and July 2005 and 2007. US DeptComm, NOAA Tech Memo NMFS-AFSC-183

Gende SM, Sigler MF (2006) Persistence of forage fish ‘hotspots’ and its association with foraging Steller sea lions(Eumetopias jubatus) in southeast Alaska. Deep-Sea Res II53:432–441

Georges JY, Bonadonna F, Guinet C (2000) Foraging habitatand diving activity of lactating Subantarctic fur seals inrelation to sea surface temperature at Amsterdam Island.Mar Ecol Prog Ser 196:291–304

Gillooly JF, Brown JH, West GB, Savage VM, Charnov EL(2001) Effects of size and temperature on metabolic rates.Science 293:2248–2251

Graham WM, Pages F, Hamner WM (2001) A physical contextfor gelatinous zooplankton aggregations: a review. Hydro-biologia 451:199–212

Gustafson EJ (1998) Quantifying landscape spatial pattern:What is the state of the art? Ecosystems 1:143–156

Halekoh U, Højsgaard S, Yan J (2006) The R package geepackfor generalized estimating equations. J Stat Software 15:1–11

Hanski I (1998) Metapopulation dynamics. Nature 396:41–49 Hargis CD, Bisonette JA, David JL (1997) Understanding

measures of landscape pattern. In: Bissonette JA (ed)Wildlife and landscape ecology: effects of pattern andscales. Springer-Verlag, New York, p 231–261

Heath RB, DeLong R, Jameson V, Bradley D, Spraker T (1997)Isoflurane anesthesia in free ranging sea lion pups. J WildlDis 33:206–210

156


Hindell MA, Burton HR, Slip DJ (1991) Foraging areas ofsouthern elephant seals, Mirounga leonia, as inferredfrom water temperature data. Aust J Mar Freshw Res 42:115–128

Holmes EE, York AE (2003) Using age structure to detectimpacts on threatened populations: a case study withSteller sea lions. Conserv Biol 17:1794–1806

Holmes EE, Fritz LW, York AE, Sweeney K (2007) Age-struc-ture modeling reveals long-term decline in the natality ofwestern Steller sea lions. Ecol Appl 17:2214–2232

Hood WR, Ono KA (1997) Variation in maternal attendance pat-terns and pup behavior in a declining population of Stellersea lions (Eumetopias jubatus). Can J Zool 75:1241–1246

Hunt GL Jr, Mehlum F, Russell RW, Irons D, Decker MB,Becker PH (1999) Physical processes, prey abundance,and the foraging ecology of seabirds. In: Adams NJ, Slo-tow RH (eds) Proc 22nd Int Ornithol Cong, Durban.BirdLife South Africa, Johannesburg, p 2040–2056

Johnston TMS, Cheriton OM, Pennington JT, Chavez FP(2009) Thin phytoplankton layer formation at eddies, fila-ments, and fronts in a coastal upwelling zone. Deep-SeaRes II 56:246–259

King JT, Gelatt TS, Pitcher KW, Pendleton GW (2007) A field-based method for estimating age in free-ranging Stellersea lions (Eumetopias jubatus) less than twenty-fourmonths of age. Mar Mamm Sci 23:262–271

Ladd C, Jahncke J, Hunt GL, Coyle KO, Stabeno PJ (2005)Hydrographic features and seabird foraging in AleutianPasses. Fish Oceanogr 14 (Suppl 1):178–195

Lander ME, Loughlin TR, Logsdon MG, VanBlaricom GR,Fadely BS, Fritz LW (2009) Regional differences in the spa-tial and temporal heterogeneity of oceanographic habitatused by Steller sea lions. Ecol Appl 19:1645–1659

Liang KY, Zeger S (1986) Longitudinal data analysis usinggeneralized linear models. Biometrika 73:13–22

Loughlin TR (1998) The Steller sea lion: a declining species.Biosphere Conserv 1:91–98

Loughlin TR, York AE (2000) An accounting of the sources ofSteller sea lion, Eumetopias jubatus, mortality. Mar FishRev 62:40–45

Loughlin TR, Sterling JT, Merrick R, Sease JL, York AE (2003)Diving behavior of immature Steller sea lions (Eumetopiasjubatus). Fish Bull 101:566–582

Lovett GM, Jones CG, Turner MG, Weathers KC (2003)Ecosystem function in heterogeneous landscapes. In:Lovett GM, Jones CG, Turner MG, Weathers KC (eds)Ecosystem function in heterogeneous landscapes. SpringerScience + Business Media, New York, p 1–4

Lunn NJ, Boyd IL, Barton T, Croxall JP (1993) Factors affect-ing the growth rate and mass at weaning of Antarctic furseals at Bird Island, South Georgia. J Mammal 74:908–919

McAllister DC, Calkins DG, Pitcher KW (2001) Underwatercapture of juvenile Steller sea lions with scuba: a narratedvideo presentation. In: Jewett SC (ed) Cold water divingfor science. Alaska Sea Grant College Program, Universityof Alaska Fairbanks, AK-SG-01-06, p 53–56

McCafferty DJ, Boyd IL, Taylor RI (1998) Diving behaviour ofAntarctic fur seal (Arctocephalus gazella) pups. Can JZool 76:513–520

McConnell BJ, Chambers C, Fedak MA (1992) Foraging ecol-ogy of southern elephant seals in relation to the bathy-metry and productivity of the Southern Ocean. Antarct Sci4:393–398

McGarigal K, Marks BJ (1995) FRAGSTATS: spatial patternanalysis program for quantifying landscape structure. GenTech Rep PNW-GTR-351, USDA Forest Service, PacificNorthwest Research Station, Portland, OR

McGarigal K, Cushman SA, Neel MC, Ene E (2002)FRAGSTATS: spatial pattern analysis program for cate-gorical maps. Available at www.umass.edu/landeco/research/fragstats/fragstats.html

McIntyre NE, Wiens JA (1999) Interactions between land-scape structure and animal behavior: the roles of hetero-geneously distributed resources and food deprivation onmovement patterns. Landscape Ecol 14:437–447

Melin SR (2002) The foraging ecology and reproduction of theCalifornia sea lion (Zalophus californianus californianus).PhD dissertation, University of Minnesota, St Paul, MN

Merrick RL (1995) The relationship of the foraging ecology ofSteller sea lions (Eumetopias jubatus) to their populationdecline in Alaska. PhD dissertation, University of Wash-ington, Seattle, WA

Merrick RL, Loughlin TR (1997) Foraging behavior of adultfemale and young-of-the-year Steller sea lions in Alaskanwaters. Can J Zool 75:776–786

Moore SE, Watkins WA, Daher MA, Davies JR, Dahlheim ME(2002) Blue whale habitat associations in the northwestPacific: analysis of remotely-sensed data using a geo-graphic information system. Oceanography 15:20–25

NMFS (National Marine Fisheries Service) (2008) Recoveryplan for the Steller sea lion, eastern and western distinctpopulation segments of Steller sea lion (Eumetopias juba-tus). NMFS, National Oceanic and Atmospheric Adminis-tration, Silver Spring, MD

Okkonen SR, Weingartner TJ, Danielson SL, Musgrave DL,Schmidt GM (2003) Satellite and hydrographic observa-tions of eddy-induced shelf-slope exchange in northwest-ern Gulf of Alaska. J Geophys Res 108:3033, doi:10.1029/2002JC001342

Pascual MA, Adkinson MD (1994) The decline of the Stellersea lion in the Northeast Pacific: demography, harvest orenvironment? Ecol Appl 4:393–403

Pitcher KW, Rehberg MJ, Pendleton GW, Raum-Suryan KL,Gelatt TS, Swain UG, Sigler MF (2005) Ontogeny of diveperformance in pup and juvenile Steller sea lions inAlaska. Can J Zool 83:1214–1231

Raum-Suryan KL, Rehberg MJ, Pendleton GW, Pitcher KW,Gelatt TS (2004) Development of dispersal, movementpatterns, and haul-out use by pup and juvenile Steller sealions (Eumetopias jubatus) in Alaska. Mar Mamm Sci 20:823–850

Ream RR, Sterling JT, Loughlin TR (2005) Oceanographic fea-tures related to northern fur seal migratory movements.Deep-Sea Res II 52:823–843

Rehberg MJ (2005) Pattern matters: changes in the organiza-tion of swimming and diving behavior by Steller sea lionjuveniles in coastal Alaska waters. MS thesis, University ofAlaska, Anchorage, AK

Rehberg MJ, Burns JM (2008) Differences in diving andswimming behavior of pup and juvenile Steller sea lions(Eumetopias jubatus) in Alaska. Can J Zool 86:539–553

Richmond JP, Burns JM, Rea LD, Mashburn KL (2005) Postna-tal ontogeny of erythropoietin and hematology in free-ranging Steller sea lions (Eumetopias jubatus). Gen CompEndocrinol 141:240–247

Richmond JP, Burns JM, Rea LD (2006) Ontogeny of totalbody oxygen stores and aerobic dive potential in Stellersea lions (Eumetopias jubatus). J Comp Physiol B 176:535–545

Roughan M, Mace AJ, Largier JL, Morgan SG, Fisher JL,Carter ML (2005) Subsurface recirculation and larvalretention in the lee of a small headland: a variation on theupwelling shadow theme. J Geophys Res 110:C10027, doi:10.1029/2005JC002898

157


Royer F, Fromentin JM, Gaspar P (2004) Association betweenbluefin tuna schools and oceanic features in the westernMediterranean. Mar Ecol Prog Ser 269:249–263

Sinclair EH, Zeppelin TK (2002) Seasonal and spatial differ-ences in diet in the western stock of Steller sea lions.J Mammal 83:973–990

Sinclair E, Loughlin TR, Pearcy W (1994) Prey selection bynorthern fur seals (Callorhinus ursinus) in the easternBering Sea. Fish Bull 92:144–156

The Open University (2001) Ocean circulation, 2nd edn. But-terworth-Heinemann, Oxford

Trillmich F, Ono KA (1991) Pinnipeds and El Niño: responsesto environmental stress. Springer-Verlag, New York

Trites AW, Miller AJ, Maschner HDG, Alexander MA andothers (2007) Bottom-up forcing and the decline of Stellersea lions (Eumetopias jubatus) in Alaska: assessing theocean climate hypothesis. Fish Oceanogr 16:46–67

Turner MG, O’Neill RV, Gardner RH, Milne BT (1989) Effectsof changing spatial scale on the analysis of landscape pat-tern. Landscape Ecol 3:153–162

Turner MG, Gardner RH, O’Neill RV (2001) Landscape ecol-ogy in theory and practice; pattern and process. Springer-Verlag, New York

Weimerskirch H, Inchausti P, Guinet C, Barbraud C (2003)Trends in bird and seal populations as indicators of a sys-tem shift in the Southern Ocean. Antarct Sci 15:249–256

Weise MJ, Harvey JT, Costa DP (in press) The role of bodysize in individual-based foraging strategies of a topmarine predator. Ecology

White EP, Brown JH (2003) The template: patterns and pro-cesses of spatial variation. In: Lovett GM, Jones CG,Turner MG, Weathers KC (eds) Ecosystem function inheterogeneous landscapes. Springer Science + BusinessMedia, New York, p 31–47

Wiens JA (1989) Spatial scaling in ecology. Funct Ecol 3:385–397

Wiens JA (2002) Riverine landscapes: taking landscape ecol-ogy into the water. Freshw Biol 47:501–515

With KA, Cadaret SJ, Davis C (1999) Movement responses topatch structure in experimental fractal landscapes. Ecol-ogy 80:1340–1353

Wolanski E, Hamner WM (1988) Topographically controlledfronts in the ocean and their biological influence. Science241:177–181

Womble JN, Sigler MF (2006) Seasonal availability of abun-dant, energy-rich prey influences the abundance and dietof a marine predator, the Steller sea lion Eumetopias juba-tus. Mar Ecol Prog Ser 325:281–293

Yan J, Fine J (2004) Estimating equations for associationstructures. Stat Med 23:859–874

York A (1994) The population decline of northern sea lions,1975–1985. Mar Mamm Sci 10:38–51

York AE, Merrick RL, Loughlin TR (1996) An analysis of theSteller sea lion metapopulation in Alaska. In: McCulloughDR (ed) Metapopulations and wildlife conservation. IslandPress, Washington, DC, p 259–292

Zeger SL, Liang KY (1986) Longitudinal data analysis for dis-crete and continuous outcomes. Biometrics 42:121–130

158

Editorial responsibility: Daniel Costa,Santa Cruz, California, USA

Submitted: March 30, 2009; Accepted: November 26, 2009Proofs received from author(s): February 3, 2010

foraging effort of juvenile steller sea lions eumetopias jubatus

Documents